The Short Form of the Interferon α/β Receptor Chain 2 Acts as a Dominant Negative for Type I Interferon Action

We have characterized the functional properties of the short form of the human interferon α/β receptor chain 2 (IFNAR2), denoted IFNAR2.1. IFNAR2.1 contains a shortened cytoplasmic domain when compared with the recently cloned full-length IFNAR2 chain (IFNAR2.2). We show that IFNα8 and IFNβ1b induce antiviral and antiproliferative activity in mouse cell transfectants expressing the human IFNAR1 chain of the receptor and induce the formation of STAT1/STAT2 dimers in IFN-stimulated response element (ISRE)-dependent gel shift assays. In contrast, coexpression of IFNAR2.1 with IFNAR1 reduces the IFN-induced antiviral, antiproliferative and ISRE-dependent gel shift binding activity conferred by IFNAR1 alone. No antiviral or antiproliferative response to IFN, nor IFN-induced ISRE-dependent gel shift binding activity, was observed when IFNAR2.1 was expressed alone in murine cells. Therefore, IFNAR2.1 acts as a dominant negative for these IFN-induced activities. Our results suggest that IFNAR2.1 represents a nonfunctional version of the full-length chain (IFNAR2.2).

Type I interferons (IFNs), 1 consisting of ␣ and ␤ subtypes, are a multifunctional cytokine family capable of inhibiting cell proliferation and viral replication, in addition to modulating cellular immune functions (1). Type I IFNs compete with one another for binding to a common multisubunit receptor present on the surface of target cells. The type I IFN receptor consists of at least two distinct subunits IFNAR1 and IFNAR2 (2)(3)(4)(5).
The IFNAR1 chain appears to be involved primarily in signal transduction (2,6,7), while the IFNAR2 chain plays a role both in ligand binding and signal transduction (3)(4)(5). The IFNAR2 chain was originally identified as a Ϸ50-kDa protein (IFNAR2.1) (3), having a truncated cytoplasmic domain due to alternative mRNA splicing (8). A 100-kDa form of the IFNAR2 chain (IFNAR2.2) has been cloned which reconstitutes biological activity in both murine (4) and human cells (5). In most human cells IFNAR2.1 is expressed at low levels relative to the full-length IFNAR2.2 chain (9). These findings call into question the biological function of IFNAR2.1.
We have examined the biological consequences of expression of the human IFNAR1 and IFNAR2.1 chains on the sensitivity of murine cells to several type I IFNs (IFN␣2, IFN␣8, and IFN␤1b). Murine L929 cells are completely unresponsive to human IFN (6,7). We previously established that the IFNAR1 subunit acts as a species-specific signal transduction component of the human type I IFN receptor complex, but it does not directly bind IFN (6,7).
In the present study, we demonstrate that IFNAR1 expression in L929 cells confers sensitivity to the antiviral, antiproliferative, and ISRE-dependent gel shift binding activities of IFN␣8 and to a lesser extent IFN␤1b. However, expression of IFNAR1 in L929 cells does not confer sensitivity to the biological activities of IFN␣2. IFNAR2.1 expression in murine cells does not confer sensitivity to any IFN subtype examined. Interestingly, cells expressing both receptor chains exhibited a markedly reduced sensitivity to the antiviral, antiproliferative, and ISRE-dependent gel shift binding activities of IFN␣8 and IFN␤1b, when compared with cells expressing only IFNAR1. These results suggest that IFNAR2.1 acts in a dominant negative manner for the induction of biological activity by several type I IFNs.
Antiviral and Antiproliferative Assays-For determining antiviral activity, cell cultures (5 ϫ 10 5 cells/25 cm 2 flask) were preincubated overnight with IFN, followed by infection with VSV (Indiana strain) for 1.5 h at 0.1 plaque-forming units (pfu) per cell. At 24 h post-infection, the virus yield in the medium was assayed by plaque formation on Vero cells (11). Assay of the antiproliferative action of IFN was performed by treating cells (1 ϫ 10 5 cells/25 cm 2 flask) with IFN. After 4 days, the cells were trypsinized and enumerated in a Coulter Counter (12,13).
Nuclear Extracts and Gel Shift Assays-Nuclei were extracted from control and IFN-treated cells (3000 IU/ml, 15 min) with buffer containing 20 mM Tris-HCl, pH 7.85, 250 mM sucrose, 0.4 M KCl, 1.1 mM MgCl 2 , 5 mM ␤-mercaptoethanol, and 0.4 mM phenylmethylsulfonyl fluoride, and extracts were frozen on dry ice and stored at Ϫ80°C (11,14,15). For gel shift analysis, the nuclear extracts were incubated with a 32 P-end-labeled ISG15 ISRE promoter probe (5Ј-GATCCATGCCTCGG-GAAGGGAAACCGAAACTGAAGCC-3Ј) (16,17) at 25°C for 30 min, and the unbound radiolabeled probe was separated from protein⅐DNA * This work was supported by a grant from Berlex Biosciences. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  complexes on 5% polyacrylamide gels. Gels were quantitated by phosphorimage autoradiography (Molecular Dynamics).

RESULTS AND DISCUSSION
Antiviral Activity of Human IFNs in Murine Transfectants-We have previously described the isolation of murine L929 cells stably expressing the cloned human IFNAR1 chain (L929R 1 cells), the 51-kDa IFNAR2.1 chain (L929R 2 cells), or both IFNAR1 and IFNAR2.1 (L929R 12 cells) (10). IFNs are defined by their ability to inhibit the replication of a wide variety of RNA and DNA viruses. Mouse L929 cells are resistant to the antiviral action of huIFN, but are highly responsive to mouse IFN (6). Expression of the human IFNAR1 (huIFNAR1) chain in L929 cells has been found to result in antiviral protection by some human type I IFNs, without affecting sensitivity to mouse IFN␣/␤ (6, 7). In the present study, we determined the effect of huIFNAR2.1 expression on IFNЈs antiviral activity in cells expressing IFNAR1. Cells were treated overnight with human IFN, then infected with VSV (at a multiplicity of infection of 0.1 pfu/cell) and the virus released into the culture medium at one day post-infection was titered by plaque assays on the indicator Vero cell line.
Expression of huIFNAR1 or huIFNAR2.1 had no effect on the ability of L929 cells to support VSV replication, since virus replication was similar in all cells lines examined (L929, L929R 1 , L929R 2 , or L929R 12 ) varying between 0.8 and 1.3 ϫ 10 8 pfu/ml. As shown in Fig. 1 even at a concentration as high as 3,000 IU/ml, IFN␣2, IFN␣8, or IFN␤1b did not markedly inhibit VSV replication in L929 cells or in L929R 2 cells. Consistent with our previous studies (6), L929R 1 cells, which only express huIFNAR1, were sensitive to the antiviral action of human IFN␣8 and to a lesser extent IFN␤1b. In contrast, coexpression of huIFNAR2.1 with huIFNAR1 in L929R 12 cells markedly reduced sensitivity to the antiviral action of human IFN␣8 or IFN␤1b. For example, IFN␣8 treatment (3000 IU/ml) induced only a 225-fold reduction in viral titer in L929R 12 cells, as compared with a Ͼ26,000-fold reduction in L929R 1 cells (Ͼ100-fold reduction in antiviral activity). A dose-dependent reduction of viral titer in L929R 1 and L929R 12 cells was observed with IFN␣8 concentrations varying from 100 to 10,000 IU/ml (Fig. 1B). However, at all concentrations the antiviral effect of IFN␣8 was markedly reduced in cells coexpressing IFNAR2.1. These results show that the expression of IFNAR2.1 reduces the sensitivity of IFNAR1-expressing cells to the antiviral action of human IFN␣8 and IFN␤1b.
Antiproliferative Activity of Human IFNs in Murine Transfectants-Besides their broad antiviral action, IFNs also inhibit cell proliferation. Since L929R 1 transfectants were highly sensitive to the antiviral action of human IFN, we extended our studies to determine if they were also sensitive to IFN's antiproliferative action. Although cells sensitive to the antiviral action of IFN are usually also sensitive to the antiproliferative action, there are instances where there is a divergence in sensitivity to these actions of IFNs (18). Thus, the various L929 transfectants were treated with human type I IFNs and cell numbers quantitated after 4 days. As shown in Fig. 2, treatment with human IFN␣8 or IFN␤1b inhibited the proliferation of L929R 1 cells which express IFNAR1 alone, or of L929R 12 cells, which coexpress IFNAR1 and IFNAR2.1. However, the antiproliferative effect in L929R 12 cells was markedly less when compared with the effect with L929R 1 cells. IFN␣2, IFN␣8, or IFN␤1b did not affect the proliferation of L929 cells or L929R 2 cells that express the IFNAR2.1 chain. IFN␣2 had no significant antiproliferative activity in any of the transfectant (L929R 1 , L929R 2 , or L929R 12 ) cell lines. Although an antiproliferative effect could be detected in L929R 1 cells at 100 IU/ml IFN␣8 (Fig. 2B) or IFN␤1b (data not shown), the maximal antiproliferative effect was observed at 10,000 IU/ml, and no further inhibition of proliferation was observed at higher IFN concentrations. Most importantly, IFN's antiproliferative effect in L929R 12 cells was reduced at all IFN concentrations when compared with the effect in L929R 1 cells. For example, the IC 50 of the antiproliferative effect of IFN␣8 was 125 IU/ml in L929R 1 transfectants as compared with an IC 50 of Ͼ30,000 IU/ml for IFN␣8 in L929R 12 cells. Thus, expression of huIF-NAR2.1 in murine cells reduced the sensitivity of cells expressing huIFNAR1 to the antiviral and antiproliferative effects of IFN␣8 and IFN␤1b.
Induction of ISGF3 Activity by Human IFNs in Murine Transfectants-An early event induced by the binding of IFNs to their receptor is specific gene induction through the activation of the STAT-containing ISGF3, which binds to the conserved ISRE present in the promoter of many IFN-responsive genes. Because of the importance of ISGF3 in IFN signaling, we determined the IFN-induced formation of the ISGF3 transcription factor in the various murine transfectants. Nuclear extracts prepared from control and IFN-treated (3,000 IU/ml, 15 min) transfectants were incubated with a labeled ISRE probe, and the resultant protein⅐DNA complexes were analyzed by the gel shift assay. Fig. 3 shows that treatment of L929R 1 cells with human IFN␣8 or IFN␤1b induced ISGF3 activity. IFN␣2 did not induce ISGF3 activity in L929R 1 cells. In contrast, the level of ISGF3 activity induced by IFN␣8 and IFN␤1b was markedly reduced in L929R 12 cells coexpressing IFNAR1 and IFNAR2.1. No ISGF3 activity was induced in L929R 2 cells by any type I IFN tested. Quantitative analysis of the autoradiograms showed that there was a Ϸ75% reduction in the formation of ISGF3 in L929R 12 transfectants exposed to either IFN␣8 or IFN␤1b when compared with ISGF3 formation in L929R 1 transfectants (Fig. 3). To confirm the presence of specific STAT proteins in the IFN-inducible protein⅐DNA complexes, we performed gel supershift assays with STAT-specific antisera and showed that the ISRE-dependent gel shift binding activity contains only STAT1 and STAT2 (Fig. 3B). These results show that IFNAR2.1 coexpression reduces ISRE-dependent gel shift binding activity in IFNAR1-expressing transfectants induced by human IFN␣8 or IFN␤1b.
General Discussion-The data presented here indicate that IFNAR1 is able to transduce signals not only for the antiviral action of some type I IFNs as shown previously (6, 7), but also for the antiproliferative action of human IFN. This presumably occurs through the interaction of the human IFNAR1 chain with components of the mouse type I IFN receptor (19), since IFNAR1 expression does not increase human IFN binding to murine cells. It is unclear why particular human type I IFNs such as IFN␣2 are unable to transduce signals through the available murine IFN receptor subunits, while IFN␣8, and to a lesser extent IFN␤1b, are able to elicit biological effects through the endogenous receptor. These results suggest that type I IFN subtypes may differentially interact with their common multisubunit receptor. For example, while IFN␣8 productively interacts with human type I IFN receptor components in a murine context, IFN␣2 does not.
In addition coexpression of IFNAR1 and IFNAR2.1 in L929 cells markedly reduced sensitivity to the biological activities of human IFN␣8 and IFN␤1b relative to cells expressing only IFNAR1, although the sensitivity to murine IFN␣/␤ was similar in all transfectants (data not shown). Therefore, IFNAR2.1 acts in a dominant negative manner for the antiviral and antiproliferative actions of human IFN␣8 and IFN␤1b, as well as for the induction of ISRE-dependent gel shift binding activity. The level of human IFNAR2.1 chain expression relative to that of the endogenous murine IFNAR2 chain in the various transfectants is unknown. Presumably murine cells do not express a short form of IFNAR2, because human IFNAR2.1 is encoded by an Alu cassette, which is a sequence present solely in primates (8). Therefore, we interpret our results to mean that human IFNAR2.1 competes with the endogenous mouse full-length IFNAR2 for interaction with type I IFNs to bind ligand and transduce biological effects.